Optimizing Critical Components to Fit into Confined Spaces Using the Adjoint Solver

Optimizing components that must fit into tight spaces can be a daunting task, even for the most experienced designer. Consider the HVAC system of a car, which supplies air to the vehicle’s cabin. Today, air conditioning is deemed standard equipment even in entry-level automobiles, so manufacturers must build it in. Its critical components – manifold ductwork — are located under the hood amid the well-planned jumble of engine, radiator, battery, transmission, and auxiliary structures. Not much room in there … and that’s just one of the complications.

Automotive designers have to fit HVAC system ductwork in increasingly confined spaces while maintaining or even improving performance.
Image courtesy Beaudaniels.com

Beyond packaging constraints, designers must optimize the overall flow rate to handle specific HVAC requirements. First, there must be enough air coming into the car; furthermore, the air flow must be uniform enough to keep back-seat passengers comfortable. Consumers, manufacturers and regulatory agencies demand that the system be energy efficient. And R&D teams insist on faster development time of components that can be manufactured as inexpensively as possible, with resulting robust performance in real-world conditions. In other words, a system that works first time, every time.

As a result, engineers must develop the right manifold pipe geometry to fit into the constrained space under the car’s hood minimizing pressure drop to ensure adequate flow distribution. Reducing pressure drop through the pipe between inlet and outlet can decrease energy consumption. Improving flow uniformity helps to ensure that air gets circulated throughout the cabin.

Design engineers once used labor-intensive trial-and-error methods to optimize ductwork. Component shapes can be extremely complex, governed by hundreds of parameters (or more). It is impossible to consider all of them. How do you make sure that you select the relevant parameters? And even if you select just one set of key design shape parameters, you still have a very large number of designs to evaluate. Consequently, this approach is not very efficient, since the design team must create geometry and mesh, run lots of tests, look at feedback, create new geometry and mesh, then do more tests, and repeat the cycle to reach a solution — maybe not the best solution, and maybe not by the product-launch deadline.

When an ANSYS automotive customer told us that they were weary of guessing geometry parameters and simulating hundreds of design points, we set about developing simulation capabilities that would automate the study of confined inner flows. I worked with Chris Hill, Chief Technologist for the ANSYS Fluids Business Unit, to develop software code that resulted in the adjoint technology. It gives fast, specific insight into finding the ideal solutions for problems such as reducing pressure drop.

Once we worked out the technical challenges and the customer was satisfied, we deployed this adjoint technology globally, because we knew that it could positively impact the design process for a lot of organizations, in a number of industries. It was a lot of work, but I am proud to be part of the development team, and the results have made it all worthwhile. (You can read more about Chris Hill’s insights on the project.)

ANSYS adjoint technology for confined inner flows recommends and even automatically implements design enhancements, morphs the mesh to a more-optimal shape, and predicts the performance improvement. You can import a STL file — which is easy to generate in CAD, mesh software, or another design tool — and define the region as a real environment. Because adjoint simulation provides a roadmap as to what regions are the most sensitive to the results you want to achieve, and which regions are not, it has a lot of power. For example, one case study shows a reduction in total pressure drop of the pipe between inlet and outlet; it also increased flow uniformity at the outlet.

In this example, the total pressure drop through a manifold duct is minimized by modifying duct geometry; at the same time, outflow distribution uniformity is maximized. The duct is required to remain within a complex bounding surface, defined by an imported mesh, while the inlet and outlet are fixed. In this case, the ANSYS adjoint solution reduced total pressure drop of the pipe between inlet and outlet by 75 percent and improved outflow uniformity by 45 percent, with 50 design iterations.

The reduction of total pressure (left vertical axis and square symbols) and outlet mass-flux mean variance (right vertical axis and triangle symbols) along with design iterations

Manifold geometry (silver) and bounding surface (green)

Manifold flow path line colored by total pressure: Disappearance of the recirculation region suggests more-efficient flow transport.

I hope you’re interested in taking advantage of this feature that is applicable to confined inner flows in cars, trucks and airplanes as well as other tight spaces. You can find more information about the adjoint solver and other advanced CFD capabilities here.

Learn some tricks and tips to get the most out of the adjoint solver in this short video and keep watching to see Part 2 !